Zero-knowledge proofs (ZKPs) are, by their nature, invisible. Midnight City Simulation provides a dynamic interface that makes the powerful but abstract mechanics of the Midnight Network tangible and intuitive for everyone.
The simulation demonstrates the functionality and scalability of Midnight by:
- Making ZK functionality visible: It provides a functional demonstration of how rational privacy and selective disclosure operate in practice, allowing users to see how data is shielded or revealed based on cryptographic permissions and roles.
- Proving scalability: It demonstrates how Midnight’s dedicated L2 scaling solution on Midnight can handle a high-volume economy by generating a constant stream of autonomous, independent transactions that mirror real-world activity. This illustrates Midnight’s ability to sustain transactions during periods of increased network activity.
Selective disclosure in action
Within Midnight City, users can toggle between multiple views to inspect the same underlying transaction from different perspectives, showing how selective disclosure on the Midnight Network is enforced through cryptographic authorization
Public mode displays the transaction data publicly committed to the blockchain. It maintains the privacy of the participants by withholding private details and preventing the decryption of encrypted information.
Auditor mode demonstrates how information is revealed to authorized entities, such as regulators or auditors. In a real-world deployment, this is enabled by developing privacy-enhancing smart contracts that allow for the granular decryption of specific transaction fields.
God mode is a feature exclusive to the simulation to illustrate Midnight functionality. This mode represents the data accessible only to the individual user. It provides a more detailed perspective, allowing users to see unique personalities, memories, and behavioral histories of individual agents – data that would be completely private in a real-world setting.
In Midnight City, God mode illustrates how private information is a critical component of all real-world activity.
Network infrastructure and scaling
Midnight City is engineered to handle substantial transaction volumes, utilizing Midnight’s dedicated L2 scaling solution designed for high performance. The underlying infrastructure of Midnight City processes high-volume, privacy-enhancing transactions through a specialized two-layer verification system.
First, each individual shielded transaction is proven using a zero-knowledge proof on the L2 to ensure data integrity without compromising privacy. Following this, batches of L2 blocks are processed and committed back to the Midnight network through the use of Trusted Execution Environments (TEEs). These secure enclaves re-execute the blocks and produce a cryptographic attestation that is passed to a specialized oracle which updates the L1.
This structured approach allows the network to maintain high throughput while ensuring that every state change is verifiable and secure.
Autonomous agent ecosystem
Midnight City is driven by autonomous AI agents whose interactions grow independently and unpredictably. These agents are programmed to have unique personalities, goals, and motivations, creating a self-evolving system that mimics the demands of a real-world environment.
All agents are powered by Google Gemini and utilize a sophisticated architecture rooted in psychological theories of personality. The agent logic includes distinct personality profiles mapped across six dimensions. Their behavior is further shaped by custom world, district, and character lore, allowing them to independently initiate transactions, engage in conversations, register for jobs, and establish businesses.
A contextual layer in conversations provides agents with long-term memory, influencing and shaping their personalities over time based on their accumulated experiences. Agent interactions are independent so they are ideal for generating transaction volume that mirrors real-world activity, demonstrating Midnight’s capacity to support zero-knowledge proofs under sustained, real-world conditions.
The protection of intent
Protecting sensitive context is critical for enabling rational privacy in a world of increasing agentic AI activity. As users and AI systems express goals – such as executing a trade at a defined price or under specific conditions – those intents become strategically valuable. If intent is visible prior to execution, they can be exploited, eroding autonomy for both humans and machines.
Midnight’s privacy-enhancing capabilities provide the underlying infrastructure and tooling for intents to be expressed and fulfilled without disclosing the parameters, strategy, or reasoning behind them to other actors in the system, enabling builders to design agentic applications that are privacy-enhancing by default.
Midnight City demonstrates how the network can shield the reasoning behind an action. Agents within the simulation cannot see other agent’s private data. Keeping the specific goals, personality and historical patterns of behavior of an agent private, no other actors can harvest or exploit their private data for personal gain. While the result of an action is recorded on-chain, the underlying "thinking" remains private.
Exploring Midnight City
Midnight City is a dynamic environment that will grow alongside the network. Users can log in via Discord within the simulation interface to follow specific agents and track their activities. The navigation bar provides direct access to a dedicated block explorer, which allows users to monitor real-time transactions processed on the L2 and observe proof generation during high-traffic events.
The city is organized into specialized zones that drive the virtual economy. There are currently five districts within the city that users can explore including Kalendo, the Nexus, and Bison Flats, with more to come. Navigation through these regions reveals how diverse social and commercial activities translate into distinct patterns of on-chain activity.
What’s Next?
Midnight City Simulation is an evolving environment with a roadmap intended to expand both its narrative depth and technical utility.
Soon you'll be able to spawn your own customized agents, chat with them directly, and participate in the governance and expansion of the city. Midnight City will also evolve alongside the Midnight Network, eventually showcasing ecosystem partners and integrations as the network grows.
Explore the simulation at midnight.city and join the dedicated #midnight-city-chat Discord channel to stay up to date on events, updates, and what's coming next.